Theoretical LENR device would be comparable to an RTG [1] because at high heat the device would melt itself. It's more about a trickle of stable power, rather than being used as a cell of a power plant. So, ignoring support infrastructure and focusing on the size of the power generator, we have:
Using the power source for the voyager probe as an example:
RTG size: 0.5m x 0.5m x 1m == 0.25 cubic meters.
RTG power: 2400 watts (thermal)
RTG power density: 9600 watts per cubic meter
Using ITER as an example for the scale of fusion power plant:
Reactor size: 800 cubic meters
Reactor power: 500,000,000 watts (thermal)
Reactor power density: 625,000 watts per cubic meter
So yeah, theoretically it would be compact, but low power density. Enough to power a space probe, but not our civilization.
[1] https://en.wikipedia.org/wiki/Radioisotope_thermoelectric_ge...